Twin-lines japh

[ambrus]

With a big thanks to Jenda for noticing this perl syntax,

($_=Just,$\="l hacker")=~(S_&&y,Jk csluath, hate Porn,),print,
($_=Just,$\="l hacker")=~($_&&y,Jk csluath, hate Porn,),print,

[blazar]

($_=Just,$\="l hacker")=~(S_&&y,Jk csluath, hate Porn,),print,
($_=Just,$\="l hacker")=~($_&&y,Jk csluath, hate Porn,),print,

++, very Abigail-ish IMHO: it neither really features data nor logic obfu, in the sense that it's easy to see where the data is and have at least a naive understanding of what the logic does, but it's surprising that is works at all that!

The japh is a single statement, composed of four comma separated expressions. The first one is

($_=Just,$\="l hacker")=~(S_&&y,Jk csluath, hate Porn,)

On the lhs $_ and $\ are assigned 'Just' and 'l hacker' respectively and $\ is returned by the comma operator in scalar context and then is bound to the expression on the rhs. The latter is equivalent to (but more funny than - thanks to a savvy choice of the order of charachters)

'S_' && tr/ Jachkls-u/a otnh erP/)

and since S_ is a true constant, the transliteration is carried on, which makes 'l hacker' into ' another', just as if 'S_' && were not there at all.

Actually, this was the most surprising part for me, that is that you can use a complex expression on the rhs of =~: in fact I most often rely on aliasing to $_ and implicit operation of matches, substitutions and transliterations on it, but retrospectively it's very perlish and obvious that an expression would be allowed to. To quote from perldoc perlop:

Read more... (1255 Bytes)

Here, the crucial point is that S_ is a (bareword) constant, hence the emphasis above. (Is this to be considered a side effect of a compile time optimization?) In fact, passing to the second line, it's just the same as the first one except for having $_ instead of S_. In which case the full expression

$_ && tr/ Jachkls-u/a otnh erP/)

is evaluated and its return value applied as pattern march, which does nothing. But as a side effect the same transliteration as above is applied, this time implicitly to $_ which makes 'Just' into ' Per' and leaves 'l hacker' untouched.

Thus, all in all, the two prints print respectively 'Just another', ' Perl hacker'.

[ambrus]

There's a current discussion on perl5porters about how /$x/ behaves specially when $x is empty (and this is true to /\Q$x/ as well). That magic is totally way crazyer than the one I'm using in this obfu (or the similar one in Re^3: A cleaner way of scoping variables; or the one you get if you combine either of these with inlined constant subroutines). My obfu is topped.

[ambrus]

P5p have decided that the feature I'm using here is a bug, and so perl 5.13.6 breaks this obfu. Try this quickily while it still works.